* tweak rx_fifo_offset post-load check to avoid possible overflow * use uint32_t rather than int32_t for rx_fifo_offset, next_packet, * don't bother updating vmstate version in intermediate patches * drop now-unneeded minimum_version_id_old entries in vmstate

* fix +/-4 error in guard on rx_fifo_offset I think patch 7 is the only one still needing review. Guest-to-host as a security boundary for those boards Unmaintained boards, so nobody IMHO should be treating * there are many lurking problems with most of our more-or-less * migration doesn't work on the stellaris board anywayīecause not all the board's devices support it, so nobody Incoming migration state), I haven't cc'd stable on it However implementing the PHY registers is more work than I wantĪlthough patch 7 fixes a CVE (code is exploitable by malicious I tested this with a minor hack to make all MII registers returnĠx24, which happens to satisfy the test image's setup code. That needs proper implementation of the MII registers in the PHY. This isn't actually sufficient to get my test image to work: I also fixed a couple of other bugs I noticed while State fields which are easier to migrate and to validate in In the first patch, and to reimplement things using simpler The principal motivation is to fix the buffer overrun noted This patchset overhauls the stellaris_enet TX and RX handling code,

` (6 more replies) 0 siblings, 7 replies 10+ messages in threadįrom: Peter Maydell 11:53 UTC ( / raw) 11:53 ` hw/net/stellaris_enet: Restructure tx_fifo code to avoid buffer overrun Peter Maydell Stellaris_enet: overhaul tx/rx, convert to vmstate All of help / color / mirror / Atom feed * stellaris_enet: overhaul tx/rx, convert to vmstate 11:53 Peter Maydell

Stellaris patch 1.9